<html><body><div style="color:#000; background-color:#fff; font-family:times new roman, new york, times, serif;font-size:12pt"><div>Salut,<br><br>Imi cer scuze ca tot pe tine de deranjez cu aceasta problema care nu te priveste, dar nu stiu la cine sa apelez si am nevoie de sfatul cuiva.<br></div><div>Multumesc pentru ajutor si scuze pentru mail-ul lung.<br></div><div><br>La cursul de SPRC s-au prezentat o <span id="misspell-5" class="">serie</span> de proiecte dintre care unul despre <span id="misspell-5" class="">Taint Analysis</span>. Mi s-a parut interesant si l-am contact pe domnul profesor responsabil pentru mai multe detalii. <br>Astazi am primit de la dansul<span id="misspell-10" class=""> niste</span> materiale de studiu <span id="misspell-11" class="">si urmeza</span> sa ma apuc de ele. Problema este ca din ultima discutie reiese ca ar trebui sa ma decid serios daca vreau sa fac asta sau nu si a<br>am doua lucruri de care imi este
teama:<br><br>1) Profesorul coordonator este de la o universitate din <span id="misspell-19" class=""><span id="misspell-23" class="">strainatate</span> si ma gandesc</span> ca ar fi mai dificil sa obtin ajutor din partea <span id="misspell-25" class="">dansului daca raman</span> blocat intr-o problema. <span id="misspell-27" class="">Ma gandesc</span> acum la <br>faptul ca va trebui sa plec de la un proiect de dimensiuni mari <span id="misspell-28" class="">si daca</span> ajung la un modul pe care nu pot sa <span id="misspell-29" class="">il inteleg</span> cum <span id="misspell-30" class="">functioneaza si</span> este vital sa <span id="misspell-31" class="">il inteleg</span>, altfel nu pot sa continui. <span id="misspell-33" class="">In acelasi</span> timp <br>nu stiu exact cat de mult ajutor se poate cere de la coordonator indiferent daca este in sala alaturata sau in alta <span id="misspell-48" class="">tara si</span> deci nu imi dau seama cat de
mult ma va deranja ca nu pot sa am o intalnire<br>directa cu dansul. Poti sa imi spui <span id="misspell-55" class="">in termeni concreti</span> cam ce neajunsuri pot sa am din cauza asta sau la ce fel de ajutor pot sa <span id="misspell-56" class="">ma astept</span> din partea profesorului coordonator ?<br><br>2) Nu pot sa evaluez cat de dificil este ceea ce isi propune proiectul. Vreau sa iti cer <span id="misspell-60" class="">tie parerea daca asa</span> ceva este fezabil (ca volum de munca si dificultate) pentru un student care este interesat de subiect,<br>dar nu vrea sa <span id="misspell-62" class="">sacrifice invatatul</span> pentru scoala. Mentionez ca sunt la C3. Includ mai jos descrierea proiectului.<br><br>" 32 TAINT ANALYSIS ON ARM (ANDROID)<br>Taint analysis is a technique to detect sophisticated memory corruption attacks. The idea is simple. We run our operating system in a modified emulator that tags all data coming from a suspect
source (such as the Internet) as 'tainted'. When such data is copied, or used as source operand in an arithmetic operation, the destination is also tainted. Likewise, taint can be cleared -- for instance, when a constant value is stored at a previously tainted address. Using taint analysis, we can detect most memory corruption attacks: essentially, we raise an alert whenever the program counter gets tainted [1]. In the last few years, papers about taint analysis have appeared in all top venues in systems and security: OSDI, SOSP, Security & Privacy, CCS, USENIX Security, NDSS, EUROSYS, etc. One of the most popular tainting system is Argos, developed at VU University [1]. urrently, we have limited our work mostly to the x86 architecture. However, with the increasing importance of ARM (for instance, in mobile devices), we now want to turn our attention to this architecture also. We believe that there will be a lot of interest for such a
system.<br>Requirements: We are looking for a real 'systems' student to develop a version of taint analyis/Argos for the ARM. Since both Argos and the Android emulator are based on QEMU, we could even target Android the project. Good coding skills are definitely required. Beyond taint analysis, the project offers opportunities for very advanced further research. Examples include tracking ROP code, attack analysis, etc.<br>The VU supervision will be handled by Prof. Herbert Bos<br>(http://www.cs.vu.nl/~herbertb/) and his team.<br>Coordonatori: Prof. Herbert Bos, (herbertb@cs.vu.nl) "</div><div><br></div><div><br></div>Multumesc. </div></body></html>